Germline mutations in the MEN1 gene encoding menin predispose to endocrine tumors mainly of the parathyroids, anterior pituitary and entero-pancreatic endocrine tissues. We have investigated the molecular basis of this tissue specific tumorigenesis from menin loss in the pathogenesis of tumors of the pancreatic islet beta-cells (insulinoma). It is possible that the cause of the tissue-specificity is due to menin-mediated regulation of one or more tissue-specific factors such as those that control differentiation during embryogenesis. Therefore, we assessed the effect of menin loss or gain on the expression of factors that are known to control beta-cell differentiation. We found that the beta-cell differentiation factor HLXB9 (MNX1) is post-transcriptionally upregulated upon menin loss. HLXB9 causes apoptosis in the presence of menin, in MIN6 insulinoma beta-cells. Thus, dysregulation of HLXB9 predicts a possible mechanism for beta-cell proliferation in insulinomas resulting from the possible blockade of the pro-apoptotic activity of HLXB9 upon menin loss. These findings advance the understanding of how a ubiquitously expressed protein such as menin controls tissue-specific tumorigenesis in the pancreas. Moreover, our data reveal the mechanisms of action of HLXB9 and its targets in beta-cells. We also showed that HLXB9 is phosphorylated by the kinase GSK-3beta, both phospho-HLXB9 and GSK-3beta are expressed in mouse and human insulinomas, and GSK-3beta inhibitors (such as lithium chloride) reduced cell proliferation and delayed cell cycle progression of mouse insulinoma cell lines. In order to understand the molecular mechanisms by which phospho-HLXB9 promotes tumorigenesis, we have identified interacting proteins and direct target genes of phospho-HLXB9 in insulinoma cells. We found that a survival factor NONO (Non-POU domain-containing octamer binding protein, also known as p54nrb, 54 kDa nuclear RNA binding protein) interacts specifically with the phospho isoform of HLXB9, explaining why phospho-HLXB9 could be pro-oncogenic. We showed that the minimal region of interaction in HLXB9 corresponds to an internal conserved region of unknown function. This region could be used to design a synthetic peptide to disrupt the endogenous HLXB9-NONO interaction and to study the impact on beta-cell proliferation. We have determined the biological consequence of clinically relevant HLXB9 and NONO mutations: two different homozygous germline mutations in HLXB9 (p.F248L and p.F272L) that were found in patients with neonatal diabetes, a condition with functional beta-cell deficiency; and two somatic heterozygous NONO mutations in endocrine-related tumors, p.H146R (parathyroid adenoma) and p.R293H (small intestine neuroendocrine tumor). In MIN6 cells, HLXB9/p.F248L mutant protein localized in the nucleus but lacked phosphorylation, and NONO/p.R293H mutant protein was structurally impaired. Thus, the absence of HLXB9 phosphorylation from a mutation associated with beta-cell loss (diabetes) and the abundant level of phospho-HLXB9 in a condition of excessive beta-cell proliferation (insulinoma) highlights the importance of modulating HLXB9 phosphorylation as a therapeutic target. Another target that we have identified by anti-phospho-HLXB9 ChIP-Seq is the c-MET inhibitor CBLB, which is downregulated by phospho-HLXB9 that would lead to upregulation of c-MET. Thus, our data propose that targeting the oncogenic receptor c-MET in insulinomas may be therapeutic. Indeed, insulinomas from the mouse models of menin-loss show activation of the oncogenic c-Met pathway (increased phospho-HLXB9, reduced Cblb and increased c-Met). Further investigations in insulinomas and other pancreatic neuroendocrine tumors will help to explore the relevance of these pathways and interactions, and the potential of c-MET inhibitor therapy.
